On Sun, Jul 29, 2007 at 09:15:50AM -0400, john wrote:
I ran again and attached the full minicom.cap file.
Thank you! I've spotted the problem.
[..first time reading vmlinuz..]
ide_readmany: sector 43929933 to 0x0011e160 ipb=32, sizeof(inode)=128 inode=125d80, raw_inode=125e00 offset into inode table block=128 00 00 00 00 00 00 00 00 3b 15 75 46 3b 15 75 46 3b 15 75 46 00 00 00 00 00 00 00 00 00 00 00 00
..
[..second time..]
ide_readmany: sector 43929933 to 0x0011e160 ipb=32, sizeof(inode)=128 inode=125d80, raw_inode=125e00 offset into inode table block=128 00 00 00 00 00 00 00 00 3b 15 75 46 3b 15 75 46 3b 15 75 46 00 00 00 00 00 00 00 00 00 00 00 00
..
[..reading initrd..]
ide_readmany: sector 43929933 to 0x0011e160 ipb=32, sizeof(inode)=128 inode=125d80, raw_inode=125e00 offset into inode table block=128 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
Wrong data. This is likely because the sector cache is not being invalidated properly. I'll have a look at the code later.
//Peter